A typical magnetic data recording device, such as a disk drive, includes a magnetic storage media, and a read/write head. The head is usually formed from two or more elongated pieces of a suitable ferromagnetic material such as a nickel iron (NiFe) alloy. The poles are joined at one end, called the yoke, and are separated by a precisely defined gap at the opposite end, called the tip. A wire coil is wrapped around the poles adjacent the yoke. During read and write operations, the head is positioned with the tip portion adjacent the disk; the coil provides a mechanism for driving magnetic flux into as well as sensing magnetic flux from the head.
In designing a disk drive it is generally desirable to increase the density of information recorded per unit area of disk media surface, as well as the data transfer rate and signal-to-noise ratio (SNR). These parameters are primarily determined by the specific electro-mechanical configuration of the head. For example, since the number of tracks per inch (TPI) available on a disk of a given size is determined by the width of the tip, it is desirable to keep the tip as narrow as possible. However, the yoke should be made as wide as possible, to encourage magnetic flux conduction between the poles and the coil. A wider yoke increases the available amplitude of the signal emanating from the head, and thus the number of bits per inch (BPI) recordable in a given track. Given these condition, it is well known that the preferred head pole configuration has a narrow tip and a wide yoke.
However, other parameters must normally be considered to achieve maximum recording density. For example, a ferromagnetic pole can be fabricated to have an induced anisotropy axis of preferred magnetic orientation, or "easy axis". When this axis is oriented perpendicular to the direction flux conduction, the flow of flux through the head can occur by the preferred process of rotation of the rest state of magnetization away from the easy axis. The preference for the material to remain magnetized in the easy axis direction is also referred to as the anisotropy field, H.sub.k.
Unfortunately, reducing the width of the pole beyond a critical value results in the dominance of "edge closure domains" which are not oriented in the preferred direction. Therefore, these edge closure domains encourage flux conduction by domain wall motion along the outer walls of the pole. While this may actually facilitate flux conduction at low frequencies, at higher frequencies the magnetic permeability, or .mu. of the pole becomes unacceptably low. Furthermore, defects in the head material can perturb the conduction of magnetic flux along the domain walls, resulting in Barkhausen noise. Thus, the pole tip cannot be made too narrow without adversely affecting head performance.
One technique for minimizing the size of the edge closure domains and thus for reducing the size of the tip is to increase the H.sub.k of the pole material. Several other techniques are known for adjusting the H.sub.k, such as adding cobalt to create an NiFeCo alloy; changing the proportion of Ni to Fe together with a change in the magneostrictive forces on the pole; or using other materials, such as cobalt zirconium (CoZr), cobalt zirconium neodymium (CoZrNd), iron nitride FeN, iron silica (FeSi), and the like.
A quite different concern arises with the yoke, however, which complicates the choice of a suitable H.sub.k. In particular, the yoke is less susceptible to Barkhausen noise, since its wider path prevents the dominance of edge closure domains, and thus the yoke usually provides sufficient flux conductivity. Rather, the problem with the yoke usually is that the permeability, .mu., of the material is inversely proportional to its H.sub.k. Having too low a .mu. limits the amount of flux conduction through the coil. Thus, it is generally believed that the H.sub.k of the pole material should be kept low in order to avoid adverse affects upon the overall head efficiency.
A further consideration is the thicknesses of the poles in the dimension parallel to the write track. Optimum pole thickness is dictated primarily by surface eddy currents, which originate from variations in the applied or detected magnetic field as different sections of the disk rotate past the head. At high frequencies, the presence of eddy currents generally limits flux conduction to the extent of the surface effect or "skin depth". Since thicker poles generally provide greater material for conduction, the thicker the pole, the better the conductivity. However, increasing the pole thickness beyond about twice the skin depth can adversely affect the pole's frequency response.
To optimize the SNR, other parameters should also be considered in determining the pole thickness. For example, consider that any given head configuration can be modelled by an equivalent electrical circuit consisting of an ideal inductor heaving a series resistance, R.sub.s, and a parallel resistance, R.sub.p. In order to maximize the available SNR, the output signal level, and hence, R.sub.p, must be as large as possible. Since R.sub.p is typically inversely proportional to the pole thickness, in order to achieve high R.sub.p, the thickness must be minimized. Hence, a compromise must typically be made between maximum frequency operation and maximum SNR when choosing the pole thickness.
A higher R.sub.p can also be obtained by decreasing the H.sub.k of the pole material, but this conflicts with the need to increase H.sub.k to encourage maximum flux conduction through the tip, as previously mentioned.
One approach to resolving some of these design tradeoffs has been to use a pole formed from multiple laminations. Such multiple laminations inherently provide the desired domain structure, since flux paths are provided in parallel by the laminated layers. They also increase the R.sub.p of the pole for a given pole thickness.
However, fabrication of laminated poles involves a more complicated and expensive fabrication process, since the layers must be formed by slower, more precise processes. For example, this process typically requires sputtering a sheet deposit and then ion milling the pole form with a mask, rather than the less expensive electroplating through a mask. Additionally, any ion milling required to form the pole geometry has the adverse effect of side-wall redeposition. This means that the longer the ion milling time, the greater the chance that redeposited material will build up sufficiently to provide a short circuit path between layers, thereby unfavorably changing the electrical and magnetic properties of the head.
Further discussion of the tradeoffs in thin head design and some suggested alternatives are discussed in the following papers.
Mallary, M., et al., "Frequency Response of Thin Film Heads with Longitudinal and Transverse Anisotropy", presented at the IEEE Conference on Magnetics in Apr., 1990.
Mallary, M., et al., "Three Dimensional Transmission Line Model for Flux Conduction in Thin Film Heads", presented at the 34th Annual Conference on Magnetism and Magnetic Materials (1989).
Mallary, M., and Smith, A. B., "Conduction of Flux at High Frequencies by a Charge Free Magnetization Distribution", IEEE Transactions on Magnetics, vol. MAG-24, p. 2374 (1988).
Mallary, M., "Conduction of Flux at High Frequencies in Permalloy Strips by Small-Angle Rotations", Journal of Applied Physics, Vol. 57, p. 3953 (1985).
What is needed is a thin film magnetic recording head which provides the desired domain structure in the tip and good flux conduction in the yoke region to provide for maximum density recording at acceptable SNRs. The head should be easily fabricated with minimal use of ion milling.